Hello all,
I am new to RNA-seq, and would like your input to see if my results are real or not. I am trying to see if there are any differentially expressed genes between knockout and wild-type strain with 3 animals for each genotype. In my gene_exp.diff output from cuffdiff the most significant p-value is 5.00E-05 and I have 58 genes with this value, and all of them have the exact same q-value of 0.0150853. I am a bit skeptical that all of these genes can have the exact same p and q values....
Could it be that they actually have a lower significant value, but cuffdiff cuts them all off at a p-value of 5.00E-05 and q-value of 0.0150853?
I have included a sample of my results below & I am using cufflinks 2.2.1
I would appreciate any insight into this phenomenon.
Cheers,
Yuka
value_1 value_2 log2(fold_change) test_stat p_value q_value significant
89.2154 58.9422 -0.597993 -2.83732 5.00E-05 0.0150853 yes
1.11605 0.637795 -0.807238 -2.87108 5.00E-05 0.0150853 yes
3.28464 2.19358 -0.582446 -2.43504 5.00E-05 0.0150853 yes
133.831 3.77205 -5.14892 -10.0845 5.00E-05 0.0150853 yes
439.124 707.344 0.687783 3.06388 5.00E-05 0.0150853 yes
46.5894 20.684 -1.17148 -4.54961 5.00E-05 0.0150853 yes
4.29346 0.487185 -3.1396 -5.51822 5.00E-05 0.0150853 yes
3.98649 2.53882 -0.650961 -2.34521 5.00E-05 0.0150853 yes
2.43337 7.40836 1.6062 3.70451 5.00E-05 0.0150853 yes
3.09364 1.61501 -0.937765 -2.6357 5.00E-05 0.0150853 yes
22.7107 34.5983 0.607328 2.75673 5.00E-05 0.0150853 yes
1.62867 0.384298 -2.0834 -3.16599 5.00E-05 0.0150853 yes
85.5877 40.7053 -1.07218 -3.79772 5.00E-05 0.0150853 yes
7.59285 4.52762 -0.74589 -3.00281 5.00E-05 0.0150853 yes
0.655639 1.15038 0.811131 3.2065 5.00E-05 0.0150853 yes
16.8931 9.04287 -0.901585 -3.10804 5.00E-05 0.0150853 yes
2.13982 1.15785 -0.886043 -2.26105 5.00E-05 0.0150853 yes
51.0117 70.4272 0.465303 2.23548 5.00E-05 0.0150853 yes
16.4426 4.60742 -1.8354 -6.30806 5.00E-05 0.0150853 yes
1.70664 3.06055 0.842634 2.63646 5.00E-05 0.0150853 yes
25898.9 11368.1 -1.1879 -2.67136 5.00E-05 0.0150853 yes
31.0799 43.4487 0.483332 2.38779 5.00E-05 0.0150853 yes
1.34841 0.854922 -0.657398 -2.31314 5.00E-05 0.0150853 yes
8.65166 2.89657 -1.57863 -6.03995 5.00E-05 0.0150853 yes
84.8254 128.85 0.603121 3.00014 5.00E-05 0.0150853 yes
7.3413 4.40212 -0.737838 -3.1716 5.00E-05 0.0150853 yes
35.9883 52.4989 0.544761 2.75034 5.00E-05 0.0150853 yes
5.09062 3.22742 -0.657457 -2.46845 5.00E-05 0.0150853 yes
3.54614 2.26065 -0.649511 -2.44616 5.00E-05 0.0150853 yes
5.88631 2.35884 -1.31929 -2.90402 5.00E-05 0.0150853 yes
1.63451 0.707828 -1.20739 -3.0934 5.00E-05 0.0150853 yes
7.81814 12.558 0.68371 2.81042 5.00E-05 0.0150853 yes
73.4685 110.3 0.586236 2.84624 5.00E-05 0.0150853 yes
3.8276 2.46574 -0.634417 -2.42448 5.00E-05 0.0150853 yes
22.13 14.2849 -0.631518 -2.55388 5.00E-05 0.0150853 yes
10.0575 5.06492 -0.989659 -3.61907 5.00E-05 0.0150853 yes
4.56879 2.70737 -0.754918 -2.5256 5.00E-05 0.0150853 yes
5.17134 3.10769 -0.734697 -2.88216 5.00E-05 0.0150853 yes
10.8373 6.79769 -0.672893 -2.30828 5.00E-05 0.0150853 yes
1.13635 0.563153 -1.0128 -2.01683 5.00E-05 0.0150853 yes
45.5072 76.8825 0.756562 3.63131 5.00E-05 0.0150853 yes
1.01017 0.112521 -3.16633 -4.40682 5.00E-05 0.0150853 yes
1.3256 0 #NAME? #NAME? 5.00E-05 0.0150853 yes
3.47888 0 #NAME? #NAME? 5.00E-05 0.0150853 yes
76.2555 33.5566 -1.18424 -4.46786 5.00E-05 0.0150853 yes
11.6164 19.7901 0.768611 3.5777 5.00E-05 0.0150853 yes
4.70657 3.21184 -0.551276 -2.49339 5.00E-05 0.0150853 yes
4.45358 1.18755 -1.90698 -5.88499 5.00E-05 0.0150853 yes
179.376 122.177 -0.554005 -2.59482 5.00E-05 0.0150853 yes
1.88054 1.23137 -0.610883 -2.60224 5.00E-05 0.0150853 yes
51.2297 34.7039 -0.561882 -2.35145 5.00E-05 0.0150853 yes
1.55802 5.58375 1.84152 4.20437 5.00E-05 0.0150853 yes
1.15313 0 #NAME? #NAME? 5.00E-05 0.0150853 yes
6.63245 12.3112 0.89236 3.63541 5.00E-05 0.0150853 yes
4.33681 7.43317 0.777344 2.41337 5.00E-05 0.0150853 yes
57.6986 11.8767 -2.2804 -4.98563 5.00E-05 0.0150853 yes
6.55491 2.75186 -1.25217 -2.86793 5.00E-05 0.0150853 yes
25.3647 16.1568 -0.650683 -2.48292 5.00E-05 0.0150853 yes
•
written
by
Yuka Takemon •
1 answer
If I remember correctly, cuffdiff uses (in recent versions like yours) a kind of permutation test (edit: sampling procedure) to assess the significance of DE. This is why it has got "bins" of qvalues and the lowest one is still not so low. I'm anyway not sure of this.
Whatever the cause of this is, it is a common situation, so it doesn't mean you did something wrong.
edit: just found a very partial comment of the authours on this topic: see the last point of the description of 2.1.0 release
Thanks for finding the 2.1.0 release... what a bummer. I guess I will have to learn Kallisto & Sleuth now.
•
written
by
Yuka Takemon •
Log in to answer this question.
Which version is this? I know there was an issue with this until one of the most recent versions.
Oops I should have included that above: I am using cufflinks/2.2.1
With cufflinks 2.2.0 I have the same problem. I didn't have it with older versions (I can't find the versions right now, sorry!). I wonder if the problem emerged when they started using the method of Simon and Anders for estimating dispersion across replicates.
I am using 2.2.1